VIP: Grass Valley Viper versus XL H1

David Taylor · February 20th, 2006, 08:06 PM

CineForm recently participated in another camera comparison shoot last week. It was mostly about shooting Grass Valley's Viper in its HDStream mode into a Wafian HR-1 D2D recorder. But we also brought along the XL H1 in its stock configuration to shoot the same material. Both cameras were shooting in 24p (24F) and were captured from their HD-SDI outputs at 4:2:2 chroma and 10 bits of precision as CineForm Intermediate files inside the HR-1.

I've attached two somewhat equivalent frame grabs for each camera - one against a matte and the other a real-world shot.

As much as I like the XL H1 as an awesome workhorse camera, see what you think about the Viper images - they are unbelievably clean (noise free) and exhibit no edge fringing. But hey, it's a $100K camera, right? And it includes a 9.2Mpixel sensor for each color channel.

Another interesting fact is that the files captured from the Viper were about 10% - 15% smaller than the XL H1 files when compressed into CineForm Intermediate. This is easily explained by the cleaner source from the Viper. CineForm Intermediate does a good job of recording noise which takes up more bits. These frames were exported out of CineForm Intermediate video sequences from the Premiere Pro timeline as bitmap files. I then compressed them lightly as JPGs to save storage space. I compared the JPG files with the bitmaps to ensure no visual artifacts were introduced in the JPG compression.

www.cineform.com/downloads-other/Car-Canon.zip
www.cineform.com/downloads-other/Car-Viper.zip
http://www.cineform.com/downloads-ot...creenCanon.zip
http://www.cineform.com/downloads-ot...creenViper.zip

Scott Billups directed and managed the test. I think there will be more written after the data is better digested.

David.

Tim Brown · February 20th, 2006, 09:59 PM

Thanks for posting these grabs.

I can hear the comments now; "The XLH1 darks are sooooooo much deeper than the Viper's!" ;-p

It is amazing, however, that the lines are blurring as much as they are. Exciting times indeed.

Thanks again.

Tim

Barlow Elton · February 21st, 2006, 12:07 AM

Oh come on, that Viper just doesn't have the MOJO. ;-)

And that Canon noise is really filmic. I just can't stand those super clean Viper images. Sheesh, looks like video. ;-)

Thanks for posting those grabs David. Please bring Cineform to FCP...pretty please?

David Taylor · February 21st, 2006, 12:54 AM

Quote:

Originally Posted by Barlow Elton

...And that Canon noise is really filmic.

When you see the moving sequence you're right - the XL H1 definitely has a not-unpleasant grainy characteristic.

Quote:

Originally Posted by Barlow Elton

Please bring Cineform to FCP...pretty please?

Well, Apple is moving to Intel processors.... Hmmmm....

Barlow Elton · February 21st, 2006, 01:25 AM

Quote:

Originally Posted by David Taylor

When you see the moving sequence you're right - the XL H1 definitely has a not-unpleasant grainy characteristic.

I was being a little tongue in cheek about all the noise talk and rationalizations for it. But it's nice to hear that it looks pretty good on film.
Can't say that I'm surprised though, I like the way my H1 renders noise, given that it's going to be there regardless.

Quote:

Well, Apple is moving to Intel processors.... Hmmmm....

Hot damn! Wait, I bet you're teasing...but one can hope. I'm not a fan of AIC, and I use DVCPRO HD just out of convenience and lack of viable options for the most part.

Isn't the main thing holding back Cineform from integration in FCP the 8 bit rendering engine that's still there? Hmmm...curiouser and curiouser.

Why the signifigance w/Intel? Must be that more processing power is necessary to do wavelet-based Cineform.

Isn't a Quad G5 powerful enough? Well, I'm sure it helps to not have to code for PowerPC platform now.

David Taylor · February 21st, 2006, 08:43 AM

CineForm Intermediate is a software-based codec heavily optimized for CPU implementations. In this case "heavily optimized" means we take advantage of behind-the-curtain techniques available within Intel processors including threading (multiple CPUs), plus assembly coding of Intel's special instruction set called SSE2. Porting the hand-optimized portions of code to another processor (G5) would be a lot of work.

Jason Rodriguez · February 21st, 2006, 12:32 PM

Just curious, what sort of glass was on the Viper? . . . I can see from the lack of CA, the purity of the specular highlights, and lack of lens flares that the lens on the Viper was definitely in a different class from the (what I assume was the) stock lens on the Canon.

Of course it also seems as though the Canon may also be more succeptable to blooming than the Viper, but again, I'm just wondering resolution wise if this was a Canon stock glass vs. Digiprime or Fujiprime test-it doesn't invalidate anything, but sort of shows you why the big boys pay so much for their glass.

Pete Tomov · February 21st, 2006, 12:45 PM

Well if I get the pick the Viper image is much better.
However,I do think that with a much better glass and a little tweaking the H1 can come pretty close.

BTW,keying greenscreen shot with H1 seems much easier that keying one shot with Viper...weird?!

EDIT: Also the lens on the H1 is designed for much higher contrast since HDV can't handle low contrast shots properly.Changing the lens and capturing via HD-SDI is going to increase the latitude a bit.

David Taylor · February 21st, 2006, 04:51 PM

Digiprime lenses were used on the Viper. I think 40mm was used inside and 20mm outside.

Simon Wyndham · March 18th, 2006, 05:33 AM

Although the H1 was using stock settings, what makes the difference for me is the amount of edge enhancement. Since the Viper has none it looks more filmic by default. In the shot of the car for instance, look at the mountain edges. The H1 looks very much more like video due to the edge enhancement.

Mike Marriage · March 18th, 2006, 08:08 AM

The Viper is clearly pulling a little more out of the shadows than the H1 and is a fair bit less noisy. I think that the Sharpening needs turning down on the H1 too. Apart from that, I could happily watch a film shot on either.

You pay a lot more for the extra 10 or 15%. Is it worth it? Depends on your budget I guess.

Simon Wyndham · March 18th, 2006, 08:48 AM

Where the shadows are concerned, that depends on what mode you put the Viper in. If you want to see just how wide a dynamic range the Viper can capture try these out for colour correcting in After Effects!

http://www.cinematography.net/Viper/viper_dpx_files.htm

Although I'm not actually sure why the Viper itself costs as much as it does. From what I gather it doesn't have half as much electronics in it as the average 2/3" camera because it captures everything RAW, and it doesn't have an on board recording device. So I do get the feeling with the Viper that you pay for the name, much like you do with Nike trainers. I could understand it if the full price included the Directors Friend, and perhaps a Venom as well. But on it's own, I'm at a loss.

One reason why RED can be made for much less perhaps?

Mike Marriage · March 18th, 2006, 11:35 AM

Quote:

Originally Posted by Simon Wyndham

Although I'm not actually sure why the Viper itself costs as much as it does.

Because people will pay it. :-)

February 20th, 2006, 08:06 PM	#1
David Taylor Major Player Join Date: Sep 2003 Location: Solana Beach, CA Posts: 853	VIP: Grass Valley Viper versus XL H1 - just for fun! CineForm recently participated in another camera comparison shoot last week. It was mostly about shooting Grass Valley's Viper in its HDStream mode into a Wafian HR-1 D2D recorder. But we also brought along the XL H1 in its stock configuration to shoot the same material. Both cameras were shooting in 24p (24F) and were captured from their HD-SDI outputs at 4:2:2 chroma and 10 bits of precision as CineForm Intermediate files inside the HR-1. I've attached two somewhat equivalent frame grabs for each camera - one against a matte and the other a real-world shot. As much as I like the XL H1 as an awesome workhorse camera, see what you think about the Viper images - they are unbelievably clean (noise free) and exhibit no edge fringing. But hey, it's a $100K camera, right? And it includes a 9.2Mpixel sensor for each color channel. Another interesting fact is that the files captured from the Viper were about 10% - 15% smaller than the XL H1 files when compressed into CineForm Intermediate. This is easily explained by the cleaner source from the Viper. CineForm Intermediate does a good job of recording noise which takes up more bits. These frames were exported out of CineForm Intermediate video sequences from the Premiere Pro timeline as bitmap files. I then compressed them lightly as JPGs to save storage space. I compared the JPG files with the bitmaps to ensure no visual artifacts were introduced in the JPG compression. www.cineform.com/downloads-other/Car-Canon.zip www.cineform.com/downloads-other/Car-Viper.zip http://www.cineform.com/downloads-ot...creenCanon.zip http://www.cineform.com/downloads-ot...creenViper.zip Scott Billups directed and managed the test. I think there will be more written after the data is better digested. David.

February 20th, 2006, 09:59 PM	#2
Tim Brown Major Player Join Date: Jan 2003 Location: Richmond, Virginia Posts: 263	Thanks for posting these grabs. I can hear the comments now; "The XLH1 darks are sooooooo much deeper than the Viper's!" ;-p It is amazing, however, that the lines are blurring as much as they are. Exciting times indeed. Thanks again. Tim

February 21st, 2006, 12:07 AM	#3
Barlow Elton Major Player Join Date: Mar 2005 Location: UT Posts: 945	Oh come on, that Viper just doesn't have the MOJO. ;-) And that Canon noise is really filmic. I just can't stand those super clean Viper images. Sheesh, looks like video. ;-) Thanks for posting those grabs David. Please bring Cineform to FCP...pretty please?

February 21st, 2006, 08:43 AM	#6
David Taylor Major Player Join Date: Sep 2003 Location: Solana Beach, CA Posts: 853	CineForm Intermediate is a software-based codec heavily optimized for CPU implementations. In this case "heavily optimized" means we take advantage of behind-the-curtain techniques available within Intel processors including threading (multiple CPUs), plus assembly coding of Intel's special instruction set called SSE2. Porting the hand-optimized portions of code to another processor (G5) would be a lot of work.

February 21st, 2006, 12:32 PM	#7
Jason Rodriguez Trustee Join Date: Mar 2003 Location: Virginia Beach, VA Posts: 1,095	Just curious, what sort of glass was on the Viper? . . . I can see from the lack of CA, the purity of the specular highlights, and lack of lens flares that the lens on the Viper was definitely in a different class from the (what I assume was the) stock lens on the Canon. Of course it also seems as though the Canon may also be more succeptable to blooming than the Viper, but again, I'm just wondering resolution wise if this was a Canon stock glass vs. Digiprime or Fujiprime test-it doesn't invalidate anything, but sort of shows you why the big boys pay so much for their glass.

February 21st, 2006, 12:45 PM	#8
Pete Tomov Regular Crew Join Date: Sep 2005 Posts: 106	Well if I get the pick the Viper image is much better. However,I do think that with a much better glass and a little tweaking the H1 can come pretty close. BTW,keying greenscreen shot with H1 seems much easier that keying one shot with Viper...weird?! EDIT: Also the lens on the H1 is designed for much higher contrast since HDV can't handle low contrast shots properly.Changing the lens and capturing via HD-SDI is going to increase the latitude a bit.

February 21st, 2006, 04:51 PM	#9
David Taylor Major Player Join Date: Sep 2003 Location: Solana Beach, CA Posts: 853	Digiprime lenses were used on the Viper. I think 40mm was used inside and 20mm outside.

March 18th, 2006, 05:33 AM	#10
Simon Wyndham Trustee Join Date: Dec 2003 Location: Malvern UK Posts: 1,931	Although the H1 was using stock settings, what makes the difference for me is the amount of edge enhancement. Since the Viper has none it looks more filmic by default. In the shot of the car for instance, look at the mountain edges. The H1 looks very much more like video due to the edge enhancement.

March 18th, 2006, 08:08 AM	#11
Mike Marriage Major Player Join Date: Sep 2005 Location: NE of London, England Posts: 788	The Viper is clearly pulling a little more out of the shadows than the H1 and is a fair bit less noisy. I think that the Sharpening needs turning down on the H1 too. Apart from that, I could happily watch a film shot on either. You pay a lot more for the extra 10 or 15%. Is it worth it? Depends on your budget I guess.

March 18th, 2006, 08:48 AM	#12
Simon Wyndham Trustee Join Date: Dec 2003 Location: Malvern UK Posts: 1,931	Where the shadows are concerned, that depends on what mode you put the Viper in. If you want to see just how wide a dynamic range the Viper can capture try these out for colour correcting in After Effects! http://www.cinematography.net/Viper/viper_dpx_files.htm Although I'm not actually sure why the Viper itself costs as much as it does. From what I gather it doesn't have half as much electronics in it as the average 2/3" camera because it captures everything RAW, and it doesn't have an on board recording device. So I do get the feeling with the Viper that you pay for the name, much like you do with Nike trainers. I could understand it if the full price included the Directors Friend, and perhaps a Venom as well. But on it's own, I'm at a loss. One reason why RED can be made for much less perhaps?